A Blended Ice Concentration Product based on Visible/infrared and Microwave

The cryosphere exists at all latitudes an in about one hundred countries. Among all the properties of the cryosphere, ice cover, ice concentration, ice thickness, and sea ice leads are among most important. The availability, consistency and accuracy of sea ice products are critical for a wide range of applications including climate monitoring and modeling, weather forecasting, shipping, fisheries, and hazard mitigation. We present an blended, optimized, all-weather ice concentration and ice cover product based on passive microwave and visible/infrared ice concentration products. The microwave data, e.g., from the Advanced Microwave Scanning Radiometer 2 (AMSR2), provides an all-weather capability but at a relatively low spatial resolution. The visible/infrared data, e.g., from the Visible Infrared Imaging Radiometer Suite (VIIRS), provides very high spatial resolution information but is limited to clear sky conditions. The combined use of ice concentration from these two products is based on uncertainty estimation of both products, and a sound blending scheme. To estimate the uncertainties, a reference ice concentration and cover from Landsat 8 Enhanced Thematic Mapper Plus (ETM+) is determined. These high spatial resolution imageries are collocated with corresponding ice products from both VIIRS and AMSR-2 with a small time difference to determine the measurement accuracy and precision of ice concentration and cover products from relatively lower spatial resolution satellite data. The measurement accuracy of each ice product is defined as the mean difference between Landsat/SAR ice concentration and these two ice concentrations, and the measurement precision as the standard deviation of the differences. The Best Linear Unbiased Estimator (BLUE) is then applied to derive the final ice concentration under clear sky conditions.